HHS Public Access Author manuscript Author Manuscript

J Lang Soc Psychol. Author manuscript; available in PMC 2016 April 27. Published in final edited form as: J Lang Soc Psychol. 2009 September ; 28(3): 281–296. doi:10.1177/0261927X09335264.

Linguistic Predictors of Mindfulness in Written Self-Disclosure Narratives Susan D. Moore and Leslie R. Brody Boston University, Massachusetts

Abstract Author Manuscript

This study investigated whether relative changes in cognitive, emotion, temporal, and selfreference word frequencies in repeated narratives predicted improvements in mindfulness skills (i.e., nonjudgmental acceptance of present-moment experiences, observing and describing present stimuli, and acting with awareness) subsequent to narrative self-disclosure. Participants wrote repeated narratives of traumatic or daily events over 3 days. Mindfulness was assessed at baseline and 4 to 8 weeks posttask. Results indicated that relative increases in cognitive processing words (among traumatic events participants and women in both conditions) and present tense words (among all participants) significantly predicted increases in nonjudgmental acceptance, describing, or overall mindfulness. Increases in present tense words appeared to partially mediate the higher mindfulness outcomes of participants writing about daily events when compared with those writing about trauma. The findings suggest that linguistic changes in self-disclosure narratives are associated with improvements in specific mindfulness skills.

Author Manuscript

Keywords mindfulness; narratives; self-disclosure; linguistic analysis

Author Manuscript

Mindfulness is commonly defined as the focusing of one’s accepting, nonjudgmental attention on present-moment experience (Kabat-Zinn, 1990, 2005). With roots in Buddhism and other Eastern meditation traditions, training in mindfulness skills is being increasingly incorporated into Western interventions for a wide range of psychological disorders (Baer, 2006). Researchers have recently developed both operational definitions and valid assessment measures for mindfulness—two elements of primary importance to scientific developments in the field. Baer, Smith, and Allen (2004) have offered a particularly thorough treatment of the multidimensional mindfulness construct and have developed a self-report measure that significantly differentiates four components of mindfulness: accepting present-moment experiences, acting with awareness (i.e., behaving with undivided attention), observing internal and external stimuli, and describing or labeling internal and external stimuli. These aspects of mindfulness have been found to differentially relate to aspects of personality and mental health. For example, subscales corresponding to acceptance of present-moment experiences, acting with awareness, and describing stimuli

Please address correspondence to Susan D. Moore, Stanford Prevention Research Center, Stanford University School of Medicine, Medical School Office Building, MC 5411, 215 Campus Drive, Stanford, CA 94305-5411; [email protected].

Moore and Brody

Page 2

Author Manuscript

were significantly negatively related to neuroticism, general psychopathology, and experiential avoidance, whereas subscales corresponding to observing and describing of stimuli were significantly positively related to emotional intelligence and significantly negatively related to alexithymia (Baer et al., 2004). Using their own unidimensional selfreport measure of mindfulness, Brown and Ryan (2003) found that more mindful individuals displayed greater concordance of implicit and explicit affect, suggesting greater selfknowledge and attunement to implicit experience. Other research has indicated that training in mindfulness, including attending to the present moment and enhancing self-awareness, is useful in the treatment of anxiety and depression (Kabat-Zinn et al., 1992; Miller, Fletcher, & Kabat-Zinn, 1995; Teasdale, Segal, & Williams, 1995).

Author Manuscript

Recently, Brody and Park (2004) have theorized that the process of mindfulness may be involved in the narrative emotional disclosure task (Pennebaker & Beall, 1986). The narrative disclosure task, in which individuals write repeated narratives about traumatic experiences while attending to their “deepest thoughts and emotions,” requires self-directed attention that may heighten awareness of internal experiences (Brody & Park, 2004) and might be expected to generate increased awareness of both cognitive and affective states. Previous research has indicated that nonemotional narrative disclosure, in which participants are instructed to write detailed descriptions of their daily lives, may also increase mindfulness (Moore, Brody, & Dierberger, in press). However, changes in the linguistic characteristics or composition of narratives that may correlate with shifts mindfulness are as yet unknown. For both theoretical and applied reasons, it would be helpful to understand if and how the language people use to describe personal experiences significantly relates to mindfulness and changes in mindfulness.

Author Manuscript Author Manuscript

Previous linguistic analyses have linked specific patterns of changes in language use to favorable outcomes, including psychologically important outcomes. Schwartz and Drotar (2004) found that relatively less frequent use of negative emotion words over time was a significant predictor of improved health-related quality of life—but only when that change occurred within the context of increased cognitive processing words. Findings on the association between positive outcomes and higher relative usage of words denoting cognitive processes (e.g., “think,” “realize,” “understand,” and other words referring to insight and causation) are quite consistent across studies (Pennebaker, 1993; Pennebaker, Colder, & Sharp, 1990; Pennebaker & Francis, 1996; Pennebaker, Mayne, & Francis, 1997). Recently, Rivkin, Gustafson, Weingarten, and Chin (2006) found that HIV+ adults whose emotional disclosure narratives reflected relative increases in the use of causation/insight words over time reported fewer negative changes in their lives brought about by having HIV (e.g., stigma, hopelessness, and negative effects on physical and social well-being). Similarly, Alvarez-Conrad, Zoellner, and Foa (2001) reported that higher levels of cognitive processing words were associated with lower postintervention anxiety among female sexual assault victims. Finally, Ullrich and Lutgendorf (2002) found that relative increases in cognitive processing words were associated with improvements in posttraumatic psychological growth (i.e., self-reported perceived benefits, such as increased personal strength and appreciation for life). Such findings on the associations between psychological outcomes and emotional and cognitive language are intriguing; it remains to be seen whether emotional and cognitive language patterns bear similar relationships to mindfulness. J Lang Soc Psychol. Author manuscript; available in PMC 2016 April 27.

Moore and Brody

Page 3

Author Manuscript

Little attention has been paid in the literature to the potential influences of temporal words— that is, verb tense—on narrative task outcomes. Research and theoretical work on mindfulness suggests that the use of present tense verbs would be most characteristic of mindfulness, defined in part as a present awareness of momentary experiences. Pennebaker et al. (1997) reported that less frequent use of past tense verbs over the course of writing three narratives was associated with higher posttask distress within a sample of bereaved adults. The authors interpreted lack of past tense language as evidence of a failure to confront past traumas and hypothesize that this failure may be an indicator of poor coping.

Author Manuscript

Limited research also exists on the relationship between self-reference words, that is, firstperson pronouns, and narrative outcomes. Some evidence suggests that self-focused attention in the form of self-references is associated with depression (Rude, Gortner, & Pennebaker, 2004; Sloan, 2005) and avoidance of stressful experiences (Boals & Klein, 2005). Although on one hand, greater mindfulness might also be expected to be associated with heightened self-references, insofar as mindfulness presumes self-awareness and the application of language to describe one’s personal experiences, a more Eastern view of mindfulness suggests that it might be associated with diminished self-references, as people increasingly focus on the present moment and the interconnections among all experiences rather than on their unique self-identity.

Author Manuscript

Few studies have examined whether patterns of relationships between linguistic variables and outcomes vary by gender, although much research has documented that women express more frequent and more intense emotions, both positive and negative, than men do across written, oral, and self-report measures (Brody & Hall, 2000). However, recent research has not found gender differences in the average frequencies of emotional or cognitive processing words in the narrative disclosure task (Epstein, Sloan, & Marx, 2005; Owen et al., 2006). On the other hand, when Epstein et al. (2005) looked more broadly at their entire sample (i.e., control and experimental conditions together), they found that women displayed higher frequencies of cognitive processing words than men. Owen, Klapow, Roth, and Tucker (2004) reported that female breast cancer patients made greater use of cognitive processing, positive emotion, and negative emotion words than male prostate cancer patients in online support group transcripts; these patterns appeared stable, as the authors did not observe gender differences in linguistic changes over time. The limited available research on gender differences in narrative disclosure word content—and particularly changes in content over the course of repeated disclosure—suggests a need for ongoing research.

Author Manuscript

In brief, the linguistic patterns over time associated with better narrative disclosure task outcomes have included more frequent expression of cognitive processing words, moderate expression of negative emotion words, and more frequent expression of past tense verbs. Although there is little existing research to guide assumptions regarding linguistic patterns and mindfulness, some ideas are intuitively appealing. Increases in emotional content are consistent with increased mindfulness in the sense that mindfulness implies openness to both positive and negative states. Similarly, increases in cognitive processing language— including words such as “think,” “know,” and “realize”—may be consistent with heightened self-awareness and acknowledgment of cognitive states. The operational definition of mindfulness, as nonjudgmental attention to the “here and now” aspects of one’s experience,

J Lang Soc Psychol. Author manuscript; available in PMC 2016 April 27.

Moore and Brody

Page 4

Author Manuscript

suggests that relative increases in use of present tense words may also be associated with increased mindfulness. Finally, it is not clear whether mindfulness will be associated with heightened or diminished references to the self, and how these changes in self-references will be related to health outcomes.

Aims of the Present Study

Author Manuscript

Our aim in the present study was to examine whether the words people use to describe traumatic and day-to-day personal experiences during the course of repeated written narrative disclosure are related to shifts in the multidimensional construct of mindfulness. We investigated changes over time in word use, that is, changes in the relative frequencies of certain linguistic categories across 3 days of narrative writing, to begin identifying the writing processes and narrative content most likely to result in heightened acceptance, awareness, observing, and describing of individuals’ moment-to-moment experiences. We hypothesized that relative increases in cognitive processing, positive and negative emotion, and present tense words in written self-disclosure narratives would be significantly associated with increases in all four components of mindfulness. Exploratory questions, given the dearth of literature to date, concerned (a) the nature of the relationships between changes in mindfulness and changes in past tense, future tense, and self-reference word use and (b) whether or not gender differences would emerge in the relationships between relative changes in word use and changes in mindfulness. Cross-sectional relationships between baseline mindfulness and types of word use were also examined.

Author Manuscript

Previous analyses using this sample had indicated that the participants who wrote about events of their day increased more over time in one component of mindfulness than did the participants who wrote about traumatic events (Moore et al., in press). The present study did not further investigate group differences in outcome. Moreover, previous work on this sample indicated that writing about traumatic events produced narratives that were significantly higher in average frequencies of positive, negative, and overall emotion words (Moore et al., in press), as well as more cognitive processing words (Moore & Brody, 2008), than writing about events of the day. Additional work using this sample indicated that women’s narratives contained higher average frequencies of positive and overall emotion words than men’s narratives as well as greater numbers of cognitive processing words, with no interactions between writing condition and gender for these variables (Moore, 2007). In contrast to the aforementioned examinations of average word frequencies across repeated narratives, the focus of the present study is on how linguistic changes across both writing conditions related to changes in mindfulness.

Author Manuscript

Method Participants Data were gathered as part of a randomized study of narrative emotional disclosure. Young adults at a large, urban university (N = 314) participated in exchange for partial course credit (for completion of Phase 1, written narrative disclosure) and a modest payment of $15 to $30 (for completion of Phase 2, the posttask assessment). Of those who completed Phase 1, 233 participants (55% female; 63% White; M age = 18.88; SD = 1.17) also completed Phase J Lang Soc Psychol. Author manuscript; available in PMC 2016 April 27.

Moore and Brody

Page 5

Author Manuscript

2 and thus comprised the final sample. A 2 (age) ×2 (gender) analysis of variance (ANOVA) revealed a significant difference in age between men (M = 19.24; SD = 1.30) and women (M = 18.59; SD = 0.95), F(1, 231) = 18.95, p < .001. Because of this confound, age was added as a covariate in subsequent analyses. Materials

Author Manuscript

Mindfulness—The Kentucky Inventory of Mindfulness Skills (KIMS; Baer et al., 2004) is a 39-item self-report inventory. It measures four proficiencies related to mindfulness, corresponding to distinct subscales: the Accept without Judgment subscale (9 items), which measures one’s tendency to apply nonjudgmental acceptance to present-moment experiences (e.g., “I criticize myself for having irrational or inappropriate emotions” [reverse scored]); the Act with Awareness subscale (10 items), which corresponds to the focusing of one’s attention on present activity (e.g., “I get completely absorbed in what I’m doing, so that all my attention is focused on it”); the Observe subscale (12 items), which measures tendencies to observe internal and external stimuli (e.g., “I pay attention to sensations, such as the wind in my hair or sun on my face”); and the Describe subscale (8 items), which refers to describing or labeling observed phenomena (e.g., “My natural tendency is to put my experiences into words”). Respondents endorse statements on a 5-point scale ranging from 1 (never or very rarely true) to 5 (very often or always true). Baer et al. (2004) reported that the inventory demonstrates adequate psychometric properties, including content validity, test–retest reliability, and internal consistency. Internal consistency of the KIMS in the present study was adequate; alpha coefficients at baseline and posttask were .84 and .85, respectively.

Author Manuscript Author Manuscript

Linguistic variables—Narratives were analyzed with Linguistic Inquiry and Word Count software (LIWC; Pennebaker, Francis, & Booth, 2001). This text analysis program evaluates narratives using a dictionary of more than 2,300 words/word stems that are grouped into 74 categories. The program provides percentages of total words used that fall into each category, thus controlling for absolute text length. Support for the LIWC’s external validity comes from agreement between independent judges and the text analysis program on the categorization of cognitive processing, emotional, past tense, and other words (Pennebaker & Francis, 1996). More recently, Kahn, Tobin, Massey, and Anderson (2007) examined the LIWC’s construct validity and concluded that it provides valid measurements of emotional content, as indicated by its ability to distinguish among narratives specifically designed to contain target emotions (e.g., the LIWC successfully detected higher percentages of positive emotion words in narratives about, or in response to, amusing events). In the present study, we examined the categories of positive emotion (e.g., “happy,” “good”), negative emotion (e.g., “hate,” “worthless”), and cognitive processing words (e.g., “think,” “realize,” “could,” “should”; this includes causation, insight, and discrepancy word subcategories). Selfreferences (first-person pronouns, e.g., “I,” “me”) and the temporal categories of present (e.g., “is,” “am”), past (e.g., “was,” “been”), and future tense words (e.g., “will”) were also examined. To investigate shifts in word usage over time, change scores were computed by subtracting word category percentages at the first writing session from those at the third session. Thus, higher change scores indicate increases in specific word categories over the course of 3 days’ writing. The procedure for computing change scores is similar to that

J Lang Soc Psychol. Author manuscript; available in PMC 2016 April 27.

Moore and Brody

Page 6

Author Manuscript

described in previous narrative disclosure studies (Pennebaker et al., 1997; Pennebaker & Francis, 1996). Procedure

Author Manuscript

The present study included two phases. In Phase 1, after providing written informed consent, participants completed baseline measures including a demographics questionnaire and the KIMS. Subsequently, individuals were randomly assigned to one of two conditions: a traumatic events writing condition or a daily events writing condition. After receiving a general overview, traumatic events participants were instructed to write about a traumatic or upsetting experience; daily events participants were instructed to write nonemotional descriptions recounting how they spent their time the previous day. Writing sessions lasted 20 minutes and took place in private rooms. Participants returned on 2 consecutive days to repeat this writing procedure; traumatic events participants continued to write about traumatic experiences, whereas daily events participants wrote descriptions of how they spent their time during the current day and how they planned to spend their time the next day. All writing instructions were provided in printed format and were adapted from Pennebaker (1994). In Phase 2, research staff e-mailed participants 4 weeks after the third and last writing session with invitations to complete a posttask version of the KIMS via a secure website. The post-task questionnaire was modified such that participants were instructed to consider “the last few weeks” when responding. Posttask assessments occurred, on average, at 4 weeks and 4 days following the last writing session. Finally, participants were debriefed as to the purposes of the study. All study procedures were approved by the Boston University Department of Psychology Internal Review Board.

Results Author Manuscript

Preliminary Analyses

Author Manuscript

One-way ANOVAs analyzing differences between completers and noncompleters (i.e., those who did not return for the follow-up assessment 4 weeks subsequent to narrative writing) indicated that there were no significant differences between those who completed both phases of the study (N = 233) and those who did not (N = 81) in total KIMS scores at baseline, average narrative length, or linguistic difference variables representing changes in positive emotion, negative emotion, past tense, and future tense words. Analyses revealed small but significant differences by completer status for changes in cognitive processing words, F(1, 312) = 6.77, p < .01, partial η2 = .02, present tense words, F(1, 312) = 4.89, p < . 03, partial η2 = .02, and self-references, F(1, 312) = 5.28, p < .02, partial η2 = .02. In each case, higher difference scores for completers (cognitive processing words: M = 0.42, SD = 1.39; present tense words: M = 1.79, SD = 2.06; self-references: M = 0.04, SD = 2.11) relative to noncompleters (cognitive processing words: M = −0.05, SD = 1.34; present tense words: M = 1.19, SD = 2.24; self-references: M = −0.58, SD = 2.08) indicate greater increases in these word frequency categories from the first to the third narrative for completers versus noncompleters. KIMS total scores, that is, the sum of all KIMS items, averaged 123.26 (SD = 16.16) at baseline and 119.41 (SD = 16.89) at posttask. Table 1 displays means and standard

J Lang Soc Psychol. Author manuscript; available in PMC 2016 April 27.

Moore and Brody

Page 7

Author Manuscript

deviations for the KIMS scale and each KIMS subscale; additional descriptive statistics (e.g., means and standard deviations by gender and condition) have been reported elsewhere (Moore et al., in press), and indicated that although there were no significant shifts in total mindfulness for either experimental or control participants, the control group increased significantly in one component of mindfulness, nonjudgmental acceptance.

Author Manuscript Author Manuscript

Total word counts for each narrative ranged from 532 to 4,713 (M = 1088.22; SD = 357.56). On average, 44.57% (SD = 5.89) of words contained in participants’ narratives were represented in the LIWC dictionary and were included in linguistic analyses. Descriptive statistics by gender and condition for linguistic category change scores are displayed in Table 2. Two-way ANOVAs (writing condition × gender, with age as a covariate) for changes in positive emotion, negative emotion, cognitive processing, and self-reference word frequencies indicated no significant main effects or interactions. A similar two-way ANOVA for present tense word change scores indicated significant main effects for condition, F(1, 228) = 65.99, p < .001, partial η2 = .22, and gender, F(1, 228) = 8.11, p < .01, partial η2 = .03, indicating that participants who wrote about daily events displayed greater increases in present tense words than participants who wrote about traumatic events, and that women displayed greater increases in present tense words than men. A two-way ANOVA for past tense change scores revealed a significant main effect for condition, F(1, 228) = 185.40, p < .001, partial η2 = .45, indicating that daily events participants decreased more in past tense words than traumatic events participants. A one-way ANOVA for future tense change scores indicated no significant difference by gender. A Mann–Whitney test indicated that daily events participants (median = 2.09) increased significantly more in future tense words than traumatic events participants (median = 0.19), U = 469.00, p < .001, r = −.74. (The Mann–Whitney nonparametric test was conducted given a significant result for Levene’s test, F(3, 229) = 35.68, p < .001, demonstrating that that future tense change score variances were unequal by condition.) We also conducted Pearson correlations to examine cross-sectional relationships among mindfulness scores at baseline and word category frequencies on the first day of writing. Across conditions, frequency of positive emotion words was significantly negatively correlated with baseline Accept without Judgment scores (r = −.14; p < .03) and significantly positively correlated with Describe scores (r = .14; p < .04). Frequency of selfreferences was negatively correlated with Observe scores (r = −.14; p < .03). All other correlations were nonsignificant at the p < .05 level. Multiple Regression Analyses

Author Manuscript

We conducted hierarchical multiple regressions to test our hypothesis that relative increases in the frequencies of cognitive, emotional, and present tense words would predict increased posttask mindfulness. A first set of regressions focused on changes in emotion and cognitive processing words as predictors of mindfulness; a second set focused on changes in present, past, and future tense words; and a third set focused on changes in self-references. In addition to examining the KIMS total scale score, each of the four KIMS subscale scores (i.e., Accept without Judgment, Act with Awareness, Observe, and Describe) were examined as outcomes in separate regressions.

J Lang Soc Psychol. Author manuscript; available in PMC 2016 April 27.

Moore and Brody

Page 8

Author Manuscript Author Manuscript

In the first set of regressions, predictor variables included age and the selected KIMS scale or subscale at baseline (entered simultaneously in Step 1). Gender, writing condition (i.e., traumatic events vs. daily events disclosure), and cognitive processing, positive emotion, and negative emotion word change scores were entered in Step 2. Dummy variables representing two-way interactions among gender, writing condition, cognitive processing, positive emotion, and negative emotion word change scores were entered in Step 3; three-way interactions among these variables were entered in Step 4. Two significant interactions emerged in this first set of regression analyses. The interaction between writing condition and change in cognitive processing words significantly predicted Accept without Judgment scores, β = −.10, t(216) = −2.02, p < .05. Simple regressions performed separately by condition revealed that within the traumatic events writing condition, relative increases in cognitive processing words from the first to the third narrative significantly predicted increased nonjudgmental acceptance from baseline to posttask, β = .18, t(156) = 2.92, p < . 01. There were no significant findings in the daily events writing condition. A second interaction in this set of analyses, between gender and change in cognitive processing words, significantly predicted KIMS Describe scores, β = .11, t(216) = 2.34, p < .02, and KIMS total scores, β = .09, t(216) = 2.02, p < .04. Simple regressions performed separately by gender indicated that among women, relative increases in cognitive processing words significantly predicted increased describing of internal and external stimuli from baseline to posttask, β = .14, t(125) = 2.33, p < .02, and also predicted increased overall mindfulness from baseline to posttask, β = .13, t(123) = 2.03, p < .05. No significant findings were observed among men.

Author Manuscript Author Manuscript

In the second set of regressions, predictor variables included age and the selected KIMS scale or subscale at baseline (entered simultaneously in Step 1). Gender, writing condition, and present, past, and future tense word change scores were entered in Step 2. Dummy variables representing two-way interactions among gender, writing condition, and word change scores were entered in Step 3; three-way interactions among these variables were entered in Step 4. A significant main effect indicated that change in present tense words significantly predicted Accept without Judgment scores, β = .13, t(216) = 2.09, p < .04, such that increases in present tense words from the first to third narrative in both conditions were associated with relative increases in nonjudgmental acceptance from baseline to posttask. A main effect also emerged for change in future tense words significantly predicting Observe scores, β = .29, t(216) = 2.43, p < .02, such that relative increases in future tense words from the first to third narrative were associated with increased observing of internal and external stimuli from baseline to posttask. Finally, changes in past tense, β = .16, t(216) = 2.31, p < . 02, and future tense, β = .27, t(216) = 2.37, p < .02, words significantly predicted Describe scores at posttask, such that relative increases in the frequencies of these word categories from the first to third narrative were associated with increased describing of internal and external stimuli. Changes in the frequencies of present, past, and future tense words did not predict Act with Awareness scores or total KIMS scores at posttask. In the third set of regressions, predictor variables included age and the selected KIMS scale or subscale at baseline (entered simultaneously in Step 1). Gender, writing condition, and self-reference change scores were entered in Step 2. Dummy variables representing two-way interactions among gender, writing condition, and self-reference change scores were entered J Lang Soc Psychol. Author manuscript; available in PMC 2016 April 27.

Moore and Brody

Page 9

Author Manuscript

in Step 3; three-way interactions among these variables were entered in Step 4. These analyses yielded no significant findings, that is, change in frequency of self-references from the first to the third day of writing did not predict changes in KIMS scale or subscale scores. Post Hoc Mediation Analysis

Author Manuscript Author Manuscript

The question emerges as to whether the change in frequency of present tense words (which was significantly higher in the daily events writing condition) mediated the significant difference in nonjudgmental acceptance between the two writing conditions found in previous work (Moore et al., in press). Four steps using regression analyses, as described by Kenny (2008), are required to show mediation. First, an initial variable (writing condition) must be significantly related to an outcome variable (posttask Accept without Judgment scores controlling for baseline scores), which regression analysis revealed it was, β = .13, p < .01. (It is important to emphasize that according to Kenny, associations—rather than significance values—are most important in demonstrating mediation; nevertheless, significance values are provided at each step for reference.) Second, the initial variable (writing condition) must be significantly related to the potential mediator (change in present tense words), which it was, β = .46, p < .001. Third, the mediator should be related to the outcome variable when controlling for the initial variable. This condition is also met, in that a regression analysis indicated that change in present tense words was associated with Accept without Judgment scores at posttask after controlling for writing condition and Accept without Judgment scores at baseline, β = .07, although this finding was nonsignificant (p < .16). These first three steps demonstrate partial mediation. A fourth step is necessary for demonstrating complete mediation in that the relationship between the initial variable (writing condition) and the outcome variable (mindfulness) should be reduced to near 0 when the influence of the mediator (change in present tense words) is controlled. In a regression controlling for change in present tense words, writing condition continued to be associated with Accept without Judgment at posttask, β = .10, p < .07. Thus, change in present tense words did not completely mediate changes in mindfulness but may have partially mediated those changes.

Discussion

Author Manuscript

Our aim in the present study was to examine the relationships between changes in the multidimensional construct of mindfulness and changes in narrative disclosure word content (including frequencies of cognitive processing and emotion words; present, past, and future tense words; and self-references) after repeated writing about traumatic or daily events. Results of multiple regression analyses confirm that relative increases in present tense words significantly predicted increased nonjudgmental acceptance of present-moment experiences from baseline to posttask. Additional analyses indicate that the significantly greater increase in frequency of present tense words by participants writing about daily events partially mediated the difference found between writing conditions in nonjudgmental acceptance, with the group writing about daily events increasing more in this component of mindfulness than the group writing about trauma.

J Lang Soc Psychol. Author manuscript; available in PMC 2016 April 27.

Moore and Brody

Page 10

Author Manuscript

The significant association between relative increases in present tense words and selfaccepting aspects of mindfulness adds to the literature on narrative disclosure and suggests that as individuals focus more on the present moment, they develop an increasingly nonjudgmental attitude toward their ongoing experiences. This is noteworthy in that it is consistent with current clinical literature encouraging individuals to focus on the present moment as a way of decreasing anxiety about either past or future imagined trauma (Linehan, 1993).

Author Manuscript

Increases in past and future tense words, tested as exploratory variables, were significantly associated with increases in observing (future tense words) and describing of present stimuli (future and past tense words). For example, increases in future tense words were associated with increased endorsement of statements such as, “I pay attention to sensations,” and “Even when I’m terribly upset, I can find a way to put it into words.” These results are intriguing and suggest that awareness and labeling of previous and future states is related to a tendency to notice internal and external events and possibly to increased planfulness. Although cause and effect are not clear here, these ideas are consistent with Hayes and his colleagues’ ideas (Fletcher & Hayes, 2005; Hayes & Wilson, 2003) about the power that descriptive language has in symbolizing and generalizing across experiences, with our capacity for language allowing, leading, and sometimes misleading us to respond to ideas about events that have not yet occurred.

Author Manuscript

Although participants in the daily events writing condition significantly decreased more in past tense words and increased more in future tense words than did participants in the traumatic events writing condition, the particular aspects of mindfulness that past and future tense linguistic variables were significantly related to (observing and describing of stimuli) did not differ significantly for the two writing conditions (Moore et al., in press). This suggests that perhaps the two groups were not sufficiently different in the patterns of these word usages to affect mindfulness outcomes.

Author Manuscript

As predicted, relative increases in the use of cognitive processing words from the first to the third narrative were associated with increased nonjudgmental acceptance from baseline to posttask, although this was true only within the traumatic events writing condition. Among women in both conditions, increases in cognitive processing words were associated with increased describing of present stimuli and with increased overall mindfulness. These findings are consistent with previous research showing that relatively more frequent use of cognitive processing words in the narrative emotional disclosure task is associated with beneficial mental health outcomes (Alvarez-Conrad et al., 2001; Ullrich & Lutgendorf, 2002). It is puzzling that significant associations between cognitive processing words and mindfulness did not emerge among all subgroups of our sample. Because women and participants in the traumatic events condition used more cognitive processing words on average than men and participants in the daily events condition did (Moore, 2007; Moore & Brody, 2008), it may be that change in these variables relates to mindfulness only when the initial frequency of word usage is at a high enough level. Cross-sectional analyses of baseline mindfulness and linguistic categories on the first day of writing indicated that the frequency of positive emotion words was negatively correlated

J Lang Soc Psychol. Author manuscript; available in PMC 2016 April 27.

Moore and Brody

Page 11

Author Manuscript

with nonjudgmental acceptance, suggesting a somewhat defensive function of positive emotion words, or alternatively, that people able to accept experiences are also able to acknowledge less positive emotions. Frequency of positive emotions was also positively correlated with tendencies to describe one’s experiences, a finding that is consistent with research indicating that mindful attention to the present moment can ameliorate depression (Miller et al., 1995; Teasdale et al., 1995). The frequency of self-references was negatively correlated with the mindfulness subscale of observing internal and external stimuli, a finding that supports a more Eastern view of mindfulness as an experience of “oneness” with the world and a lessening in importance of individual self-identity. Each of these correlations, though significant, was weak (coefficients

Linguistic Predictors of Mindfulness in Written Self-Disclosure Narratives.

This study investigated whether relative changes in cognitive, emotion, temporal, and self-reference word frequencies in repeated narratives predicted...
123KB Sizes 0 Downloads 9 Views